Howto import an oracle dump in an different tablespace

Question

You've got a couple of issues here.

Firstly, the different versions of Oracle you're using is the reason for the table statistics error - I had the same issue when some of our Oracle 10g Databases got upgraded to Release 2, and some were still on Release 1 and I was swapping .DMP files between them.

The solution that worked for me was to use the same version of exp and imp tools to do the exporting and importing on the different Database instances. This was easiest to do by using the same PC (or Oracle Server) to issue all of the exporting and importing commands.

Secondly, I suspect you're getting the ORA-00959: tablespace 'A_TBLSPACE' does not exist because you're trying to import a .DMP file from a full-blown Oracle Database into the 10g Express Edition (XE) Database, which, by default, creates a single, predefined tablespace called USERS for you.

If that's the case, then you'll need to do the following..

1. With your .DMP file, create a SQL file containing the structure (Tables):

   imp <xe_username>/<password>@XE file=<filename.dmp> indexfile=index.sql full=y

2. Open the indexfile (index.sql) in a text editor that can do find and replace over an entire file, and issue the following find and replace statements IN ORDER (ignore the single quotes.. '):

   Find: 'REM<space>' Replace: <nothing>

   Find: '"<source_tablespace>"' Replace: '"USERS"'

   Find: '...' Replace: 'REM ...'

   Find: 'CONNECT' Replace: 'REM CONNECT'

3. Save the indexfile, then run it against your Oracle Express Edition account (I find it's best to create a new, blank XE user account - or drop and recreate if I'm refreshing):

   sqlplus <xe_username>/<password>@XE @index.sql

4. Finally run the same .DMP file you created the indexfile with against the same account to import the data, stored procedures, views etc:

   imp <xe_username>/<password>@XE file=<filename.dmp> fromuser=<original_username> touser=<xe_username> ignore=y

You may get pages of Oracle errors when trying to create certain objects such as Database Jobs as Oracle will try to use the same Database Identifier, which will most likely fail as you're on a different Database.

If you're using Oracle 10g and datapump, you can use the REMAP_TABLESPACE clause. example:

REMAP_TABLESPACE=A_TBLSPACE:NEW_TABLESPACE_GOES_HERE

For me this work ok (Oracle Database 10g Express Edition Release 10.2.0.1.0):

impdp B/B full=Y dumpfile=DUMP.dmp REMAP_TABLESPACE=OLD_TABLESPACE:USERS

But for new restore you need new tablespace

P.S. Maybe useful http://www.oracle-base.com/articles/10g/OracleDataPump10g.php

What version of Oracle are you using? If its 10g or greater, you should look at using Data Pump instead of import/export anyway. I'm not 100% sure if it can handle this scenario, but I would expect it could.

Data Pump is the replacement for exp/imp for 10g and above. It works very similar to exp/imp, except its (supposedly, I don't use it since I'm stuck in 9i land) better.

Here is the Data Pump docs

The problem has to do with the CLOB columns. It seems that the imp tool cannot rewrite the create statement to use another tablespace.

Source: http://asktom.oracle.com/pls/asktom/f?p=100:11:0::::P11_QUESTION_ID:66890284723848

The solution is: Create the schema by hand in the correct tablespace. If you do not have a script to create the schema, you can create it by using the indexfile= of the imp tool.

You do have to disable all constraints your self, the oracle imp tool will not disable them.

After that you can import the data with the following command:

imp b/*** file=oracledump.dmp log=import.log fromuser=a touser=b statistics=none ignore=y

Note: I still needed the statistics=none due to other errors.

extra info about the data pump

As of Oracle 10 the import/export is improved: the data pump tool ([http://www.oracle-base.com/articles/10g/OracleDataPump10g.php][1])

Using this to re-import the data into a new tablespace:

1. First create a directory for the temporary dump:

   CREATE OR REPLACE DIRECTORY tempdump AS '/temp/tempdump/';
   GRANT READ, WRITE ON DIRECTORY tempdump TO a;

2. Export:

   expdp a/* schemas=a directory=tempdump dumpfile=adump.dmp logfile=adump.log

3. Import:

   impdp b/* directory=tempdump dumpfile=adump.dmp logfile=bdump.log REMAP_SCHEMA=a:b

Note: the dump files are stored and read from the server disk, not from the local (client) disk

my solution is to use GSAR utility to replace tablespace name in the DUMP file. When you do replce, make sure that the size of the dump file unchanged by adding spaces. E.g.

gsar -f -s"TSDAT_OV101" -r"USERS      " rm_schema.dump rm_schema.n.dump
gsar -f -s"TABLESPACE """USERS      """ ENABLE STORAGE IN ROW CHUNK 8192 RETENTION" -r"                                                                   " rm_schema.n1.dump rm_schema.n.dump
gsar -f -s"TABLESPACE """USERS      """ LOGGING" -r"                                  " rm_schema.n1.dump rm_schema.n.dump
gsar -f -s"TABLESPACE """USERS      """ " -r"                             " rm_schema.n.dump rm_schema.n1.dump

I wanna improve for two users both in different tablespaces on different servers (databases)

1. First create a directories for the temporary dump for both servers (databases):

server #1:

CREATE OR REPLACE DIRECTORY tempdump AS '/temp/old_datapump/';
GRANT READ, WRITE ON DIRECTORY tempdump TO old_user;

server #2:

CREATE OR REPLACE DIRECTORY tempdump AS '/temp/new_datapump/';
GRANT READ, WRITE ON DIRECTORY tempdump TO new_user;

2. Export (server #1):

expdp tables=old_user.table directory=tempdump dumpfile=adump.dmp logfile=adump.log

3. Import (server #2):

impdp directory=tempdump dumpfile=adump_table.dmp logfile=bdump_table.log
REMAP_TABLESPACE=old_tablespace:new_tablespace REMAP_SCHEMA=old_user:new_user